A comparison of PAM50 intrinsic subtyping with immunohistochemistry and clinical prognostic factors in tamoxifen-treated estrogen receptor-positive breast cancer

Abstract

Purpose: To compare clinical, immunohistochemical (IHC), and gene expression models of prognosis applicable to formalin-fixed, paraffin-embedded blocks in a large series of estrogen receptor (ER)-positive breast cancers from patients uniformly treated with adjuvant tamoxifen.

Experimental design: Quantitative real-time reverse transcription-PCR (qRT-PCR) assays for 50 genes identifying intrinsic breast cancer subtypes were completed on 786 specimens linked to clinical (median follow-up, 11.7 years) and IHC [ER, progesterone receptor (PR), HER2, and Ki67] data. Performance of predefined intrinsic subtype and risk-of-relapse scores was assessed using multivariable Cox models and Kaplan-Meier analysis. Harrell's C-index was used to compare fixed models trained in independent data sets, including proliferation signatures.

Results: Despite clinical ER positivity, 10% of cases were assigned to nonluminal subtypes. qRT-PCR signatures for proliferation genes gave more prognostic information than clinical assays for hormone receptors or Ki67. In Cox models incorporating standard prognostic variables, hazard ratios for breast cancer disease-specific survival over the first 5 years of follow-up, relative to the most common luminal A subtype, are 1.99 [95% confidence interval (CI), 1.09-3.64] for luminal B, 3.65 (95% CI, 1.64-8.16) for HER2-enriched subtype, and 17.71 (95% CI, 1.71-183.33) for the basal-like subtype. For node-negative disease, PAM50 qRT-PCR-based risk assignment weighted for tumor size and proliferation identifies a group with >95% 10-year survival without chemotherapy. In node-positive disease, PAM50-based prognostic models were also superior.

Conclusion: The PAM50 gene expression test for intrinsic biological subtype can be applied to large series of formalin-fixed, paraffin-embedded breast cancers, and gives more prognostic information than clinical factors and IHC using standard cut points.

©2010 AACR.

Figures

Figure 1

Kaplan-Meier survival analysis of intrinsic subtype (panels A and B) and Risk of Relapse score (ROR-S, panels C and D), as determined by PAM50 gene expression measurement by qRT-PCR performed on paraffin blocks from women with invasive breast carcinoma, treated with adjuvant tamoxifen. The number of events and total number of patients in each group is shown beside each curve’s description. RFS = relapse-free survival (panels A and C). DSS = breast cancer disease-specific survival (panels B and D; excludes two cases with unknown cause of death).

Figure 2

C-index estimates of relapse free (RFS) and disease-specific (DSS) survival for different measures of hormone receptors and proliferation. DCC = dextran-coated ligand binding assay. IHC = % positive nuclei by immunohistochemistry. The Luminal and Proliferation measures are the means of normalized qRT-PCR values across 8 and 11 signature genes respectively, as described in Supplementary Methods. P-values were estimated from 1000 bootstrap samples. Single asterisk (*) designates significant improvement (p

Figure 3

Comparison of prognostic classifiers in node negative subjects. The C-index is used to compare accuracy of the prognostic classifiers (panel A and Supplemental Table S5). * designates significant improvement (p Adjuvant!) and ** to the IHC-T model. Taking the best-performing model, ROR-PT values are related to actual 10 year event probabilities using a Cox proportional hazard model (panel B, dotted lines are 95% CI). Kaplan-Meier survival analysis of the size and proliferation weighted Risk of Relapse (ROR-PT) assignments are presented in panels C and D, and comparable information provided by a model of IHC subtype and tumor size is shown in panels E and F. RFS = relapse-free survival. DSS = breast cancer disease-specific survival (excludes two cases with unknown cause of death)

Figure 4

Comparison of prognostic classifiers in node positive subjects. A) C-index comparison of the accuracy of prognostic classifiers as described in Figure 3. B) Cox proportional hazard model relating the best performing model (ROR-T) to actual 10 year event probabilities. C) Kaplan-Meier survival analysis of the size weighted Risk of Relapse (ROR-T) assignments for RFS and (in panel D) DFS. E) and F) comparable information as provided by a model of IHC subtype and tumor size.